Time Scale Of Adaptation Research Articles

Specificity and timescales of cortical adaptation as inferences about natural movie statistics.

Adaptation is a phenomenological umbrella term under which a variety of temporal contextual effects are grouped. Previous models have shown that some aspects of visual adaptation reflect optimal processing of dynamic visual inputs, suggesting that adaptation should be tuned to the properties of natural visual inputs. However, the link between natural dynamic inputs and adaptation is poorly understood. Here, we extend a previously developed Bayesian modeling framework for spatial contextual effects to the temporal domain. The model learns temporal statistical regularities of natural movies and links these statistics to adaptation in primary visual cortex via divisive normalization, a ubiquitous neural computation. In particular, the model divisively normalizes the present visual input by the past visual inputs only to the degree that these are inferred to be statistically dependent. We show that this flexible form of normalization reproduces classical findings on how brief adaptation affects neuronal selectivity. Furthermore, prior knowledge acquired by the Bayesian model from natural movies can be modified by prolonged exposure to novel visual stimuli. We show that this updating can explain classical results on contrast adaptation. We also simulate the recent finding that adaptation maintains population homeostasis, namely, a balanced level of activity across a population of neurons with different orientation preferences. Consistent with previous disparate observations, our work further clarifies the influence of stimulus-specific and neuronal-specific normalization signals in adaptation.

The time scale of adaptation in tonal sequence processing by the mouse auditory midbrain neurons.

The time course of poststimulatory adaptation of the inferior colliculus central nucleus (ICC) of CBB6F1 hybrid mice to sound sequences, specifically, series of four tonal stimuli presented at intervals of 0, 2, 4, 10, 20, 50, 100, 200, 500, 700, 1000, and 1500 ms were studied. Assessment of the adaptation of the entire neuronal population have shown that, at an interstimulus interval of 0-200 ms, the response to the first tone in a series is significantly stronger than those to the second to fourth tones, the strengths of the latter three responses not differing significantly from one another. If the interstimulus interval is 500 ms or longer, the response to none of the tones in a series differs significantly in strength from the others. The role of adaptation of midbrain neurons to the grouping of components of bioacoustic stimuli is discussed.

Hydrodynamic modelling and characterisation of a shallow fluvial lake: a study on the Superior Lake of Mantua

This paper presents a numerical modelling framework developed to simulate circulations and to generally characterise the hydrodynamics of the Superior Lake of Mantua, a shallow fluvial lake in Northern Italy. Such eutrophied basin is characterised by low winds, reduced discharges during the summer and by the presence of large lotus flower (<em>Nelumbo</em> <em>nucifera</em>) meadows, all contributing to water stagnation. A hydrodynamic numerical model was built to understand how physical drivers shape basic circulation dynamics, selecting appropriate methodologies for the lake. These include a 3D code to reproduce the interaction between wind and through-flowing current, a fetch-dependent wind stress model, a porous media approach for canopy flow resistance and the consideration of wave-current interaction. The model allowed to estimate the circulation modes and water residence time distributions under identified typical ordinary, storm and drought conditions, the hydrodynamic influence of the newly-opened secondary outlet of the lake, the surface wave parameters, their influence on circulations and the bottom stress they originate, and the adaptation time scales of circulations to storm events. Some probable effects of the obtained hydrodynamic characteristics of the Superior Lake of Mantua on its biochemical processes are also introduced.

Robust Brain-Machine Interface Design Using Optimal Feedback Control Modeling and Adaptive Point Process Filtering.

Much progress has been made in brain-machine interfaces (BMI) using decoders such as Kalman filters and finding their parameters with closed-loop decoder adaptation (CLDA). However, current decoders do not model the spikes directly, and hence may limit the processing time-scale of BMI control and adaptation. Moreover, while specialized CLDA techniques for intention estimation and assisted training exist, a unified and systematic CLDA framework that generalizes across different setups is lacking. Here we develop a novel closed-loop BMI training architecture that allows for processing, control, and adaptation using spike events, enables robust control and extends to various tasks. Moreover, we develop a unified control-theoretic CLDA framework within which intention estimation, assisted training, and adaptation are performed. The architecture incorporates an infinite-horizon optimal feedback-control (OFC) model of the brain’s behavior in closed-loop BMI control, and a point process model of spikes. The OFC model infers the user’s motor intention during CLDA—a process termed intention estimation. OFC is also used to design an autonomous and dynamic assisted training technique. The point process model allows for neural processing, control and decoder adaptation with every spike event and at a faster time-scale than current decoders; it also enables dynamic spike-event-based parameter adaptation unlike current CLDA methods that use batch-based adaptation on much slower adaptation time-scales. We conducted closed-loop experiments in a non-human primate over tens of days to dissociate the effects of these novel CLDA components. The OFC intention estimation improved BMI performance compared with current intention estimation techniques. OFC assisted training allowed the subject to consistently achieve proficient control. Spike-event-based adaptation resulted in faster and more consistent performance convergence compared with batch-based methods, and was robust to parameter initialization. Finally, the architecture extended control to tasks beyond those used for CLDA training. These results have significant implications towards the development of clinically-viable neuroprosthetics.

Experimental studies of bone mechanoadaptation: bridging in vitro and in vivo studies with multiscale systems.

Despite advancements in technology and science over the last century, the mechanisms underlying Wolff's law-bone structure adaptation in response to physical stimuli-remain poorly understood, limiting the ability to effectively treat and prevent skeletal diseases. A challenge to overcome in the study of the underlying mechanisms of this principle is the multiscale nature of mechanoadaptation. While there exist in silico systems that are capable of studying across these scales, experimental studies are typically limited to interpretation at a single dimension or time point. For instance, studies of single-cell responses to defined physical stimuli offer only a limited prediction of the whole bone response, while overlapping pathways or compensatory mechanisms complicate the ability to isolate critical targets in a whole animal model. Thus, there exists a need to develop experimental systems capable of bridging traditional experimental approaches and informing existing multiscale theoretical models. The purpose of this article is to review the process of mechanoadaptation and inherent challenges in studying its underlying mechanisms, discuss the limitations of traditional experimental systems in capturing the many facets of this process and highlight three multiscale experimental systems which bridge traditional approaches and cover relatively understudied time and length scales in bone adaptation.

Neural Substrates Related to Motor Memory with Multiple Timescales in Sensorimotor Adaptation.

Recent computational and behavioral studies suggest that motor adaptation results from the update of multiple memories with different timescales. Here, we designed a model-based functional magnetic resonance imaging (fMRI) experiment in which subjects adapted to two opposing visuomotor rotations. A computational model of motor adaptation with multiple memories was fitted to the behavioral data to generate time-varying regressors of brain activity. We identified regional specificity to timescales: in particular, the activity in the inferior parietal region and in the anterior-medial cerebellum was associated with memories for intermediate and long timescales, respectively. A sparse singular value decomposition analysis of variability in specificities to timescales over the brain identified four components, two fast, one middle, and one slow, each associated with different brain networks. Finally, a multivariate decoding analysis showed that activity patterns in the anterior-medial cerebellum progressively represented the two rotations. Our results support the existence of brain regions associated with multiple timescales in adaptation and a role of the cerebellum in storing multiple internal models.

Analytical approach to an integrate-and-fire model with spike-triggered adaptation.

The calculation of the steady-state probability density for multidimensional stochastic systems that do not obey detailed balance is a difficult problem. Here we present the analytical derivation of the stationary joint and various marginal probability densities for a stochastic neuron model with adaptation current. Our approach assumes weak noise but is valid for arbitrary adaptation strength and time scale. The theory predicts several effects of adaptation on the statistics of the membrane potential of a tonically firing neuron: (i) a membrane potential distribution with a convex shape, (ii) a strongly increased probability of hyperpolarized membrane potentials induced by strong and fast adaptation, and (iii) a maximized variability associated with the adaptation current at a finite adaptation time scale.

Simultaneous single-step one-shot optimization with unsteady PDEs

The single-step one-shot method has proven to be very efficient for PDE-constrained optimization where the partial differential equation (PDE) is solved by an iterative fixed point solver. In this approach, the simulation and optimization tasks are performed simultaneously in a single iteration. If the PDE is unsteady, finding an appropriate fixed point iteration is non-trivial. In this paper, we provide a framework that makes the single-step one-shot method applicable for unsteady PDEs that are solved by classical time-marching schemes. The one-shot method is applied to an optimal control problem with unsteady incompressible Navier–Stokes equations that are solved by an industry standard simulation code. With the Van-der-Pol oscillator as a generic model problem, the modified simulation scheme is further improved using adaptive time scales. Finally, numerical results for the advection–diffusion equation are presented.

Three timescales in prism adaptation.

It has been proposed that motor adaptation depends on at least two learning systems, one that learns fast but with poor retention and another that learns slowly but with better retention (Smith MA, Ghazizadeh A, Shadmehr R. PLoS Biol 4: e179, 2006). This two-state model has been shown to account for a range of behavior in the force field adaptation task. In the present study, we examined whether such a two-state model could also account for behavior arising from adaptation to a prismatic displacement of the visual field. We first confirmed that an "adaptation rebound," a critical prediction of the two-state model, occurred when visual feedback was deprived after an adaptation-extinction episode. We then examined the speed of decay of the prism aftereffect (without any visual feedback) after repetitions of 30, 150, and 500 trials of prism exposure. The speed of decay decreased with the number of exposure trials, a phenomenon that was best explained by assuming an "ultraslow" system, in addition to the fast and slow systems. Finally, we compared retention of aftereffects 24 h after 150 or 500 trials of exposure: retention was significantly greater after 500 than 150 trials. This difference in retention could not be explained by the two-state model but was well explained by the three-state model as arising from the difference in the amount of adaptation of the "ultraslow process." These results suggest that there are not only fast and slow systems but also an ultraslow learning system in prism adaptation that is activated by prolonged prism exposure of 150-500 trials.

Future Arctic climate changes: Adaptation and mitigation time scales

AbstractThe climate in the Arctic is changing faster than in midlatitudes. This is shown by increased temperatures, loss of summer sea ice, earlier snow melt, impacts on ecosystems, and increased economic access. Arctic sea ice volume has decreased by 75% since the 1980s. Long‐lasting global anthropogenic forcing from carbon dioxide has increased over the previous decades and is anticipated to increase over the next decades. Temperature increases in response to greenhouse gases are amplified in the Arctic through feedback processes associated with shifts in albedo, ocean and land heat storage, and near‐surface longwave radiation fluxes. Thus, for the next few decades out to 2040, continuing environmental changes in the Arctic are very likely, and the appropriate response is to plan for adaptation to these changes. For example, it is very likely that the Arctic Ocean will become seasonally nearly sea ice free before 2050 and possibly within a decade or two, which in turn will further increase Arctic temperatures, economic access, and ecological shifts. Mitigation becomes an important option to reduce potential Arctic impacts in the second half of the 21st century. Using the most recent set of climate model projections (CMIP5), multimodel mean temperature projections show an Arctic‐wide end of century increase of +13°C in late fall and +5°C in late spring for a business‐as‐usual emission scenario (RCP8.5) in contrast to +7°C in late fall and +3°C in late spring if civilization follows a mitigation scenario (RCP4.5). Such temperature increases demonstrate the heightened sensitivity of the Arctic to greenhouse gas forcing.

Dynamic behavior analysis of fractional-order Hindmarsh–Rose neuronal model

Previous experimental work has shown that the firing rate of multiple time-scales of adaptation for single rat neocortical pyramidal neurons is consistent with fractional-order differentiation, and the fractional-order neuronal models depict the firing rate of neurons more verifiably than other models do. For this reason, the dynamic characteristics of the fractional-order Hindmarsh-Rose (HR) neuronal model were here investigated. The results showed several obvious differences in dynamic characteristic between the fractional-order HR neuronal model and an integer-ordered model. First, the fractional-order HR neuronal model displayed different firing modes (chaotic firing and periodic firing) as the fractional order changed when other parameters remained the same as in the integer-order model. However, only one firing mode is displayed in integer-order models with the same parameters. The fractional order is the key to determining the firing mode. Second, the Hopf bifurcation point of this fractional-order model, from the resting state to periodic firing, was found to be larger than that of the integer-order model. Third, for the state of periodically firing of fractional-order and integer-order HR neuron model, the firing frequency of the fractional-order neuronal model was greater than that of the integer-order model, and when the fractional order of the model decreased, the firing frequency increased.

Network Bursting Dynamics in Excitatory Cortical Neuron Cultures Results from the Combination of Different Adaptive Mechanism

In the brain, synchronization among cells of an assembly is a common phenomenon, and thought to be functionally relevant. Here we used an in vitro experimental model of cell assemblies, cortical cultures, combined with numerical simulations of a spiking neural network (SNN) to investigate how and why spontaneous synchronization occurs. In order to deal with excitation only, we pharmacologically blocked GABAAergic transmission using bicuculline. Synchronous events in cortical cultures tend to involve almost every cell and to display relatively constant durations. We have thus named these “network spikes” (NS). The inter-NS-intervals (INSIs) proved to be a more interesting phenomenon. In most cortical cultures NSs typically come in series or bursts (“bursts of NSs”, BNS), with short (∼1 s) INSIs and separated by long silent intervals (tens of s), which leads to bimodal INSI distributions. This suggests that a facilitating mechanism is at work, presumably short-term synaptic facilitation, as well as two fatigue mechanisms: one with a short timescale, presumably short-term synaptic depression, and another one with a longer timescale, presumably cellular adaptation. We thus incorporated these three mechanisms into the SNN, which, indeed, produced realistic BNSs. Next, we systematically varied the recurrent excitation for various adaptation timescales. Strong excitability led to frequent, quasi-periodic BNSs (CV∼0), and weak excitability led to rare BNSs, approaching a Poisson process (CV∼1). Experimental cultures appear to operate within an intermediate weakly-synchronized regime (CV∼0.5), with an adaptation timescale in the 2–8 s range, and well described by a Poisson-with-refractory-period model. Taken together, our results demonstrate that the INSI statistics are indeed informative: they allowed us to infer the mechanisms at work, and many parameters that we cannot access experimentally.

Does centennial morphodynamic evolution lead to higher channel efficiency in San Pablo Bay, California?

Abstract Measured bathymetries on 30 year interval over the past 150 years show that San Pablo Bay experienced periods of considerable deposition followed by periods of net erosion. However, the main channel in San Pablo Bay has continuously narrowed. The underlying mechanisms and consequences of this tidal channel evolution are not well understood. The central question of this study is whether tidal channels evolve towards a geometry that leads to more efficient hydraulic conveyance and sediment throughput. We applied a hydrodynamic process-based, numerical model (Delft3D), which was run on 5 San Pablo Bay bathymetries measured between 1856 and 1983. Model results shows increasing energy dissipation levels for lower water flows leading to an approximately 15% lower efficiency in 1983 compared to 1856. During the same period the relative seaward sediment throughput through the San Pablo Bay main channel increased by 10%. A probable explanation is that San Pablo Bay is still affected by the excessive historic sediment supply. Sea level rise and Delta surface water area variations over 150 years have limited effect on the model results. With expected lower sediment concentrations in the watershed and less impact of wind waves due to erosion of the shallow flats, it is possible that energy dissipations levels will decrease again in future decades. Our study suggests that the morphodynamic adaptation time scale to excessive variations in sediment supply to estuaries may be on the order of centuries.

Effects of neuronal adaptation currents on network-based spike rate oscillations

Local field potential in-vivo recordings often show oscillations with frequencies ranging from < 1 Hz to 100 Hz. The underlying mechanism could change across frequency bands but likely involves network interactions such as recurrent synaptic excitation-inhibition loops. Particularly fast rhythmic activity in the beta and gamma range is thought to be caused by synaptic inhibition [1]. Although these oscillations primarily depend on synaptic properties, their frequency is significantly influenced by the passive membrane characteristics of single cells [2]. In addition, neuronal membranes typically contain slowly deactivating voltage-dependent as well as calcium-activated potassium channels both which mediate spike rate adaptation. Here we investigate how the dynamics of such neuronal adaptation currents contribute to synaptically generated spike rate oscillations and resonance properties in recurrent networks of excitatory and inhibitory neurons. Based on a network of sparsely delay-coupled adapting spiking model neurons we take a mean-field approach using the Fokker-Planck equation to analyze oscillatory network activity. In the limit of slow adaptation timescales we obtain population spike rates and membrane potential distributions [3]. To allow for a linear stability analysis we derive a low-dimensional system of ordinary differential equations from the Fokker-Planck mean-field model. This system effectively describes the activity of the network while retaining the features of the spiking neurons (i.e. the model parameters). For constant external input we find that fast oscillations stabilize if recurrent inhibition is slower than excitation. The oscillation frequency increases with increasing inhibitory synaptic efficacy and decreasing inhibitory delay. Neuronal adaptation facilitates such network-based oscillations for fast synaptic inhibition and leads to decreased frequencies (see Figure ​Figure1).1). Additionally we characterize chaotic spike rate dynamics, which is produced by the interplay of spike-dependent adaptation and slow synaptic inhibition. For oscillatory external input, adaptation currents lead to amplification of a narrow band of frequencies and cause phase advances for low frequencies in addition to phase delays at higher frequencies. Our results highlight the role of intrinsic neuronal adaptation properties in determining the collective rhythms originating from network interactions. Figure 1 Influence of synaptic inhibition and neuronal adaptation on network-based oscillations. A-C: Existence of oscillatory network states (OSC) and corresponding frequencies f as a function of the strength and timescale of synaptic inhibition for networks ...

Modelling cerebellar Purkinje cells with simple neuron models of the threshold type

The Purkinje cells of the cerebellum are responsible for processing the inputs necessary for coordinating movement. In the cerebellum, cells receive external input from only two sources, climbing fibres and mossy fibres [1]. Mossy fibres excite granule cells whose axons, called parallel fibres, enervate Purkinje cells. The integration of inputs from tens of thousands of individual parallel fibres synaptically connected to a single Purkinje cell causes it to emit a 'simple spike' discharge. This is a single spike with the waveform of a typical action potential. In addition to simple spikes, cerebellar Purkinje cells can produce a second kind of electro-physiological response called a complex spike only after activation by climbing fibre inputs. The complex spike consists of an initial spike followed by sub-threshold oscillations in the membrane potential which last for tens of milliseconds. This dual responsiveness of Purkinje cells presents a challenging modelling problem. Modelling efforts have usually relied on complicated conductance based models with many compartments. On the other hand, recent results [2,3] have shown that in the case of regular cortical neurons, certain simple neuron models which feature adaptation are capable of reproducing the behaviour of real cortical neurons and of detailed Hodgkin-Huxley type models of the same to a high degree of accuracy. Two notable examples of these recent models are the adaptive exponential integrate and fire model of Brette and Gerstner [2] and the multiple adaptation time-scales (MAT) model of Kobayashi et al. [3] which is a simple integrate and fire equipped with multiple scales of adaptation in the voltage threshold. Here we evaluate how effectively the rich dynamics of cerebellar Purkinje cells can be modelled using specially designed modifications of simple neuron models of this type. We use recently developed neuron model optimisation tools for this purpose [4] and benchmark the model performance against detailed multi-compartmental conductance based models in NEURON and Genesis[5].

Drift and Diffusion in Movement Adaptation to Space-Time Constraints

Recent studies have shown more than one time scale of change in the movement dynamics of practice. Here, we decompose the drift and diffusion dynamics in adaptation to performing discrete aiming movements with different space-time constraints. Participants performed aiming movements on a graphics drawing board to a point target at 5 different space-time weightings on the task outcome. The drift was stronger the shorter the time constraint whereas noise was U-shaped across the space-time conditions. The drift and diffusion of adaptation in discrete aiming movements varied as a function of the space-time constraints on performance outcome and the spatial, temporal, or space-time measure of performance outcome. The findings support the postulation that the time scale of movement adaptation is task dependent.

Design and Analysis of Closed-Loop Decoder Adaptation Algorithms for Brain-Machine Interfaces

Closed-loop decoder adaptation (CLDA) is an emerging paradigm for achieving rapid performance improvements in online brain-machine interface (BMI) operation. Designing an effective CLDA algorithm requires making multiple important decisions, including choosing the timescale of adaptation, selecting which decoder parameters to adapt, crafting the corresponding update rules, and designing CLDA parameters. These design choices, combined with the specific settings of CLDA parameters, will directly affect the algorithm's ability to make decoder parameters converge to values that optimize performance. In this article, we present a general framework for the design and analysis of CLDA algorithms and support our results with experimental data of two monkeys performing a BMI task. First, we analyze and compare existing CLDA algorithms to highlight the importance of four critical design elements: the adaptation timescale, selective parameter adaptation, smooth decoder updates, and intuitive CLDA parameters. Second, we introduce mathematical convergence analysis using measures such as mean-squared error and KL divergence as a useful paradigm for evaluating the convergence properties of a prototype CLDA algorithm before experimental testing. By applying these measures to an existing CLDA algorithm, we demonstrate that our convergence analysis is an effective analytical tool that can ultimately inform and improve the design of CLDA algorithms.

Adaptation to Speed in Macaque Middle Temporal and Medial Superior Temporal Areas

The response of a sensory neuron to an unchanging stimulus typically adapts, showing decreases in response gain that are accompanied by changes in the shape of tuning curves. It remains unclear whether these changes arise purely due to spike rate adaptation within single neurons or whether they are dependent on network interactions between neurons. Further, it is unclear how the timescales of neural and perceptual adaptation are related. To examine this issue, we compared speed tuning of middle temporal (MT) and medial superior temporal neurons in macaque visual cortex after adaptation to two different reference speeds. For 75% of speed-tuned units, adaptation caused significant changes in tuning that could be explained equally well as lateral shifts, vertical gain changes, or both. These tuning changes occurred rapidly, as both neuronal firing rate and Fano factor showed no evidence of changing beyond the first 500 ms after motion onset, and the magnitude of tuning curve changes showed no difference between trials with adaptation durations shorter or longer than 1 s. Importantly, the magnitude of tuning shifts was correlated with the transient-sustained index, which measures a well characterized form of rapid response adaptation in MT, and is likely associated with changes at the level of neuronal networks. Tuning curves changed in a manner that increased neuronal sensitivity around the adapting speed, consistent with improvements in human and macaque psychophysical performance that we observed over the first several hundred ms of adaptation.

Distributed cerebellar plasticity implements adaptable gain control in a manipulation task: a closed-loop robotic simulation

Adaptable gain regulation is at the core of the forward controller operation performed by the cerebro-cerebellar loops and it allows the intensity of motor acts to be finely tuned in a predictive manner. In order to learn and store information about body-object dynamics and to generate an internal model of movement, the cerebellum is thought to employ long-term synaptic plasticity. LTD at the PF-PC synapse has classically been assumed to subserve this function (Marr, 1969). However, this plasticity alone cannot account for the broad dynamic ranges and time scales of cerebellar adaptation. We therefore tested the role of plasticity distributed over multiple synaptic sites (Hansel et al., 2001; Gao et al., 2012) by generating an analog cerebellar model embedded into a control loop connected to a robotic simulator. The robot used a three-joint arm and performed repetitive fast manipulations with different masses along an 8-shape trajectory. In accordance with biological evidence, the cerebellum model was endowed with both LTD and LTP at the PF-PC, MF-DCN and PC-DCN synapses. This resulted in a network scheme whose effectiveness was extended considerably compared to one including just PF-PC synaptic plasticity. Indeed, the system including distributed plasticity reliably self-adapted to manipulate different masses and to learn the arm-object dynamics over a time course that included fast learning and consolidation, along the lines of what has been observed in behavioral tests. In particular, PF-PC plasticity operated as a time correlator between the actual input state and the system error, while MF-DCN and PC-DCN plasticity played a key role in generating the gain controller. This model suggests that distributed synaptic plasticity allows generation of the complex learning properties of the cerebellum. The incorporation of further plasticity mechanisms and of spiking signal processing will allow this concept to be extended in a more realistic computational scenario.

How adaptation shapes spike rate oscillations in recurrent neuronal networks

Neural mass signals from in-vivo recordings often show oscillations with frequencies ranging from <1 to 100 Hz. Fast rhythmic activity in the beta and gamma range can be generated by network-based mechanisms such as recurrent synaptic excitation-inhibition loops. Slower oscillations might instead depend on neuronal adaptation currents whose timescales range from tens of milliseconds to seconds. Here we investigate how the dynamics of such adaptation currents contribute to spike rate oscillations and resonance properties in recurrent networks of excitatory and inhibitory neurons. Based on a network of sparsely coupled spiking model neurons with two types of adaptation current and conductance-based synapses with heterogeneous strengths and delays we use a mean-field approach to analyze oscillatory network activity. For constant external input, we find that spike-triggered adaptation currents provide a mechanism to generate slow oscillations over a wide range of adaptation timescales as long as recurrent synaptic excitation is sufficiently strong. Faster rhythms occur when recurrent inhibition is slower than excitation and oscillation frequency increases with the strength of inhibition. Adaptation facilitates such network-based oscillations for fast synaptic inhibition and leads to decreased frequencies. For oscillatory external input, adaptation currents amplify a narrow band of frequencies and cause phase advances for low frequencies in addition to phase delays at higher frequencies. Our results therefore identify the different key roles of neuronal adaptation dynamics for rhythmogenesis and selective signal propagation in recurrent networks.